Financial Analysis with ChatGPT's Super CFO GPT

Separating the wheat

We’ve all used rule-based systems, OCR tools, and Google search in our work, and they’ve been helpful. But what the latest generation of AI can do is truly mind-blowing.

Today, I wanted to see if a ChatGPT GPT could deliver meaningful financial analysis. So I put the “Super CFO – Finance, Accounting, Tax & Legal” GPT agent to the test. I asked it to analyze a publicly traded company—not a household name, but one I know well. Specifically, I asked it to review the financial statements and give me a sense of the company’s overall health.

Within seconds, the Super CFO delivered a solid overview, incorporating both stock performance and insights drawn directly from the latest filings. At first, I noticed it seemed to accept the company’s own explanations and justifications a bit too readily.

But then I asked: “What’s keeping this business afloat?”And the answer floored me.

The specialized GPT Agent read "separated the wheat from the chaff" of the financial statements and correctly identified the key factor sustaining the company—an insight that I suspect even many employees there don't realize. It was a moment that made me pause and realize just how powerful this tool could be. Next, I asked about a major year-over-year change in a financial line item (which was the result of an accounting write-off but a layperson wouldn't catch it), and the Super CFO GPT nailed that explanation too.

Chat GPT Super CFO GPT summary

Sure, a trained finance professional could reach the same conclusion if he paid attention. But how many of us really maintain such attention and do it quickly? And how often do we have the time to do this for every company we review? What used to take several hours can now be reduced to a few minutes—complete with intelligent, contextual insights.

Of course, I still recommend a “trust but verify” approach when using AI. But as a starting point for financial statement review, this is significant.

I also tested the same prompts using standard ChatGPT (without the GPTs), and the results weren’t nearly as strong. The Super CFO GPT not only gave a better overall review, but surfaced conclusions that the base model didn't know to emphasize .

There’s definitely a learning curve—prompt structure matters, and some trial and error is expected—but I highly recommend trying it out. Start with a company you know well so you can better gauge the accuracy and insight of the responses.

This is just the beginning of what’s possible.

P.S. I also tested this scenario with a couple of other models:

Following Ruben Hassid’s recommendation, I ran it through Perplexity’s Chat. The results were solid, but overall not as strong as what I saw with the Super CFO GPT.

I also tried it on GenSpark, and that chat actually nailed it — delivering sharp, spot-on insights.